Over the past ten years, the popularity of Internet use among adolescents has dramatically increased; 93% of teens between the ages 12–17 go online in the U.S., as do 93% of Japanese adolescents [1
]. China had 256 million adolescent Internet users as of January 2014, accounting for 71.8% of the overall number of adolescents [2
]. Given the swift increase in numbers of teenage users, it is not surprising that the question has been raised of whether the advantages of adolescent Internet use outweigh the disadvantages. Internet-based programs can provide opportunities for interactivity and active participation that might not otherwise be available to adolescents. There is evidence that computer game practice improves the spatial performance and iconic and visual attention skills of adolescents [3
]. However, excessive Internet use may lead to grey matter atrophy in the brain [4
], negatively affecting concentration and memory, as well as the ability to make decisions and set goals [5
]. In addition, heavy Internet use may cause psychological disorders, such as Internet addiction, depression, and anxiety [6
Studies have found that excessive Internet use is strongly associated with depressive symptoms, and that adolescents involved with problematic Internet use are vulnerable to psychological disorders, such as depression [7
]. Conversely, depression in adolescents are more likely to lead to Internet use problems [11
], which is also supported by Brunet’s longitudinal study [12
]. However, the nature and causal directionality of the relationship between problematic Internet use and depression are unclear.
Internet addiction and other problematic Internet use behaviors may have significant influence on the sleep-wake schedule, leading to insomnia and other sleep disturbances [13
]. In previous studies, heavy Internet use was found to be associated with insomnia [14
], and increased time spent on the Internet led to the significant disturbance of sleep [15
]. Other previous studies have found evidence that time spent on digital game-playing and problematic Internet use are associated with sleep delays, irregular sleeping patterns and excessive daytime sleepiness [15
], which in turn are associated with increased waking-time tiredness [18
]. One psychophysical mechanism that may help to explain the negative impact of problematic Internet use on sleeping habits may be that night-time computer usage leads to a state of high arousal, thus interfering with the calming processes that are necessary for sleep [19
Sleep disturbances are highly interrelated with a depressive status [20
]. To be diagnosed with major depression, an individual must have at least five of nine criteria for depressive symptoms, and disordered sleep is one of the nine symptoms. More than two-thirds of depressed children and adolescents suffer sleep-onset or sleep-maintenance problems [20
], even though the manifestation of sleep disturbance is not necessary for a diagnosis of major depression. Researchers have suggested that the relationship between sleep disturbance and depression is bidirectional [24
], and that the two conditions could feed back on each other to mutually maintain their existence [25
]. Longitudinal research has demonstrated that sleep disturbance is associated with an increased risk of developing depression [26
], and adolescents reporting sleep disturbances may display depressive symptoms within one year [28
]. Depression may also lead to sleep disturbance in depressed children by way of disturbing circadian regulation, maintaining a negative state, and reducing both regular exposure to bright light and social activities [29
To date, despite the relationships among problematic Internet use, depression and sleep disturbance have been explored by numerous studies, relatively few studies have examined whether problematic Internet use and depression have differential effects on sleep disturbances among adolescents. Hence, we conducted a cross-sectional study to fill the gap and explore the associations between problematic Internet use, depression and their differential effects on sleep disturbance. We predicted that: (1) problematic Internet use would be associated with higher levels of depression and sleep disturbance, and depression would be associated with sleep disturbance; and (2) problematic Internet use and depression would have differential effects on sleep disturbance.
3. Results and Discussions
Among the 1661 high school students, 51.8% were male and 48.2% were female, and the average age was 14.53, with a range of 12 to 18 years, and 96.9% of sample ranging from 13 to 16 years. The mean number of hours of Internet use per week was 6.21 h. About 18% of the students went online almost every day, and 9.6% of the students usually used the Internet for more than 5 h at a time. The mean IAT score was 36.91, with 17.2% of students being categorized as engaging in problematic Internet use. The mean CESD score was 9.69, with 54.4% of students identified as having depression symptoms, and the mean PSQI score was 5.16, with 40.0% of the students being classified as having sleep disturbance (See Table 1
shows the correlations between problematic Internet use and the two main outcome variables, depressive symptoms and sleep disturbance. The IAT score was significantly associated with depressive symptoms and sleep disturbance. The correlation between depressive symptoms and sleep disturbance was highly and statistically significant.
The associations of problematic Internet use, depressive symptoms and sleep disturbance with gender, school performance, family factors and reported negative effects of Internet use are reported in Table 3
. The t
-test analyses found the reported negative effects of Internet use on health, school performance and family relationships to be individually associated with IAT, PSQI and CESD scores. For males, lower school scores, coming from a non-intact family and high level of father’s education were associated with problematic Internet use. For females, high school grade level, lower school scores, coming from a non-intact family and low level of father’s education were factors associated with higher CESD scores. For females, high school grade, coming from a non-intact family, low level of father’s education and high level of mother’s education were factors associated with higher PSQI scores.
As Table 4
shows, model I presents the mediational model of depressive symptoms as the mediator between problematic Internet use and sleep disturbance (Figure 1
a). Problematic Internet use was found to be a significant predictor of sleep disturbance (β
= 0.048, p
< 0.001) and depressive symptoms (β
= 0.146, p
< 0.001), and depressive symptoms significantly predicted sleep disturbance (β
= 0.08, p
< 0.001). When the mediator (depressive symptoms) was controlled, problematic Internet use also significantly predicted sleep disturbance (β
= 0.014, p
= 0.001). This suggested that depression was a partial mediator between problematic Internet use and sleep disturbance, with the mediator responsible for 70.6% (0.146 × 0.232/0.048) of the whole effect. The Sobel test confirmed that depression was a significant mediator (Z
= 12.70, p
< 0.001). In model II, problematic Internet use was a mediator between depression and sleep disturbance (Figure 1
b). Depression was associated significantly with sleep disturbance (β
= 0.242, p
< 0.001) and problematic Internet use (β
= 0.717, p
< 0.001), and problematic Internet use was associated significantly with sleep disturbance (β
= 0.014, p
= 0.001). When problematic Internet use was taken into account, depression was significantly associated with sleep disturbance (β
= 0.232, p
< 0.001), and problematic Internet use was a partial mediator accounting for 4.1% (0.717 × 0.014/0.242) the whole effect. The Sobel test indicated that the problematic Internet use was a significant mediator (Z
= 3.39, p
There have been rapid increases in Internet use in society in recent years, and the potentially harmful effects of excessive Internet use on the health of individuals, and of youth in particular, have been described by many studies. However, the effects of problematic Internet use and depression on sleep disturbance are not well understood. To our knowledge, this study is the first to begin to elucidate the inter-relationship of problematic Internet use and depression on the effects of sleep disturbance among Chinese adolescents, and its underlying mechanisms.
In our study, the prevalence of problematic Internet use was 17.2% among adolescents, with 40.0% of adolescents suffering from sleep disturbance, and up to 51.4% of adolescents had depressive symptoms. These results were highly similar to a study in Hong Kong, China [37
]. The prevalences in our sample were higher than those reported in previous studies [38
]; it might be possible that we adopted different methods and criteria in data collection, and moreover, adolescents were in a special growth phase, mentally and physically unstable and undergoing great changes.
Although the Internet positively affects our daily lives, our findings suggest that problematic Internet use had negative effects on the physical health, academic performance and family relationships of adolescents, similar to a study in Taiwan [41
], which found that the Internet negatively influenced many aspects of student’s lives. Adolescents who were problematic Internet users perceived significantly stronger negative Internet influence on daily routines, school performance, and parental relationships than the non-dependents. These results showed the importance of proper supervision and monitoring of a child’s Internet use by parents and teachers.
With regard to the relationship between problematic Internet use, depressive symptoms and sleep disturbance, the current study found that problematic Internet use was positively associated with the level of depressive symptoms and sleep disturbance, as indicated by bivariate analyses, which supported our first prediction. The total scores on the Internet addiction test were correlated strongly with both the depressive symptom scores and sleep disturbance. Additionally, depressive symptoms were highly associated with sleep disturbance. A prior study showed that moderate/high Internet addiction might be applied to deal with depressive symptoms, but excessive Internet use could lead to further depressive symptoms [42
]. Depressive adolescents also suffer more easily from problematic Internet use [11
]. In addition, students with problematic Internet use may hide their vulnerable and negative feelings, such as depression, to escape from the negative or stressful events caused by excessive Internet use [43
]. In the same way, previous reports indicate that Internet addiction has a close association with sleep disturbance [15
]. Their results show that the prevalence of sleep problems is higher among Internet-addicted students, and excessive Internet use could affect our health indirectly through lack of sleep. The relation between depression and sleep disturbance is consistent with previous studies findings that depression is the main factor affecting sleep, and insomnia could increase the risk of depression and also reflect the depressive symptoms [21
Our study also extended previous research by revealing the relationships between problematic Internet use, depression and sleep disturbance, showing that problematic Internet use and depression had differential effects on sleep disturbance, which supports prediction 2. In model I, we found that when depression was a mediator, problematic Internet use had negative effects on sleep disturbance, and depression partially mediated the association between problematic Internet use and sleep disturbance, with the mediator responsible for 70.6% of the whole effect. That is to say, there was about 70.6% of the effect of problematic Internet use on sleep disturbance going though depression indirectly and the direct effect of problematic Internet use on sleep disturbance was 29.4%. In model II, depression and sleep disturbance were partially mediated by problematic Internet use, and the mediating effect of problematic internet use was responsible for 4.1% of whole effect, indicating that approximately 95.9% of the effect of depression on sleep disturbance was direct. Comparison of the results of two models showed that depression exerted a stronger mediating effect than problematic Internet use, and also indicated that depression was closely associated with sleep disturbance. These suggested that depression would play a more important role than problematic Internet use when considered as a mediator. Sobel tests of the two models showed that the Z-score in model I in which depression was a mediator was higher than in model II where problematic Internet use as a mediator, which meant that the accuracy of model I was higher than model II, further confirming the conclusion that depression exerted a stronger mediating effect than problematic Internet use. Therefore, we concluded that problematic Internet use and depression partially mediated the relationship of sleep disturbance with these two variables, and depression as a mediator exerted stronger effect for sleep disturbance than problematic Internet use.
Several limitations should be noted with regard to this study. First, because it was a cross-sectional study, our results were not able clearly to indicate causal directionality, i.e., we could not say whether problematic Internet use or depressive symptoms precede the development of sleep disturbance. Longitudinal research is needed to address this problem, and help us better understand the inter-relationship and underlying mechanisms connecting Internet overuse and depressive symptoms with sleep quality. Second, the data were reported by the students themselves, which might result in self-report bias and even undermine authenticity of data, though our investigators emphasized that the questionnaires must be finished truthfully. Third, although the standardized scales were adopted as screening tools to assess psychological problems and poor sleep quality, these measurements could not be equivalent to clinical diagnoses, therefore, more clinical diagnosis by qualified psychiatrists are needed in future studies. Fourth, other factors, such as urbancity or race, were been taken into consideration in our study. Students who are from different areas and races may have different individual chacteristics. In addition, our measurement of Internet use also had its limitations, as it did not specify the time of day of Internet use. It is likely that Internet use at night is more detrimental to the patterns and quality of adolescent sleep than daytime Internet use, as discussed above. Finally, the subjects recruited only from the junior high schools in a southern Chinese city, so one should be cautious in generalizing our findings to Chinese adolescents.